# open your kerberos session# run the yarn session clustercd$FLINK_HOME
HADOOP_CLASSPATH="`hadoop classpath`" bin/yarn-session.sh -nm "My flink experiment"# note the last log message it should indicate which address to connect to to access the flink UI and open tunnel to it e.g.:# ssh -L 8888:an-worker1103.eqiad.wmnet:42409 stat1004.eqiad.wmnet##Run the jobcd$FLINK_HOME
HADOOP_CLASSPATH="`hadoop classpath`" ./bin/flink run --python ../api_actions.py --pyarch venv.zip -pyexec venv.zip/bin/python